This is an open working note. Values and definitions here are working values, they are published so the reasoning is inspectable, not because they are settled.
Status: open. The dimension is under test and may be cut.
What continuity is trying to detect
The difference between an organisation that runs a programme and one that responds to events.
This is a real distinction and a consequential one. A reactive organisation can produce enormous volumes of completed work, all of it in the fortnight after an incident, none of it in the eleven weeks either side. Volume cannot tell those two organisations apart. Something else has to.
The primary measure
For each declared standing practice, compare expected operating opportunities with eligible completed dispositions across the minimum observation window. Continuity is the share of expected periods or events in which the practice met its declared cadence and disposition requirement.
A monthly access review completed in each of six expected periods has 100% cadence adherence. Six reviews performed in one month and none in the other five do not. An event-driven practice uses qualifying events rather than calendar months, and a quarterly practice is not forced into a monthly clock.
This replaces the earlier proposed fixed floor of n completed actions per month. That design rewarded volume, treated unlike practices as comparable and could be gamed with trivial work. Expected cadence is still imperfect, but it tests the operating promise the programme actually made.
The secondary measure, and what it revealed
Share of completed work carrying no originating correlation. Work with a correlation origin exists because a signal fired. Work without one exists because a review was due, a programme required it, a person raised it manually, or its provenance is incomplete. The absence of a correlation alone therefore does not establish proactive work. A scheduled-programme identifier and expected-due record are required for that classification.
Across most of the historical demonstration window this reads close to zero, because the original seeded work derives from correlations.
That result was expected, and the interesting thing is what it means. It is not a measurement failure. It is a finding about the historical seed: it models reactive work far more heavily than scheduled programme work. IO Coach™ can now declare a standing programme, playbook, population, cadence and owner, and generate scheduled work after approval. That capability does not retroactively rebalance the existing history, so continuity should be demonstrated on newly elapsed IO Coach™ periods rather than claimed from the seed.
A measure that exposes a gap in the product before it measures anything about a customer is doing useful work even while it is unvalidated.
The condition for keeping it
If neither measure separates on real data, the dimension gets cut.
Five dimensions that work are better than six with one hollow. A dimension retained because it was designed rather than because it discriminates is exactly the kind of thing that accumulates in maturity models until nobody can defend what the score is made of.
What would settle it
Execution history from organisations with genuinely different operating rhythms, at least one running a scheduled programme and at least one working purely reactively, over a window long enough to contain several periods. Six months minimum, and preferably across different practice cadences, since a daily practice and a quarterly control will not behave the same way.
Until that exists, continuity may be computed and displayed as an experimental measure but does not do load-bearing work in the canonical level gates. The current maturity specification uses cadence adherence at Level 2 only after the minimum observation window has closed.
Industry relationship
Institutionalisation and continual operation are established maturity concepts, including in C2M2 and ISO management systems. IO's cadence-adherence calculation is its own operational test of that idea. It does not reproduce a C2M2 MIL or an ISO conformity result.